This CDMAS study uses blinded, simultaneous interviews, to assess the ability of relatives or other surrogates to predict the treatment preferences of patients in hypothetical clinical situations. The interviews are being assessed for their accuracy.